| The existence of doubling, i.e. the multiple morphosyntactic manifestation of a single semantic property/operation in natural language, has formed a long-standing problem for compositionality. In order to account for doubling two main approaches have been proposed: (i) absorption mechanisms, where the two morphosyntactic markers are both semanticallyinterpretable and somehow merge into one operator, thus yielding no iterative readings; (ii) marking mechanisms where not every morphosyntactic marker of a semantic property/operator is semantically interpretable. I argue that this second approach is worth exploring, since it is able to yield proper readings without formulating additional interpretation mechanisms. The question however is why natural language should exhibit semantically vacuous material in the first place. In my research I especially focus on the relation between redundancy and syntactic categories. I explore the hypothesis that the existence of(semantic) redundant material functions as a cue for the analysis of syntactic categories: no syntactic category F can exist in a particular language if the language does not exhibit any doubling effects with respect to F. Redundancy is then the cause of forming syntactic categories, i.e. categories that trigger syntactic operations. Although an advantage of such a flexible approach of syntactic categories is that there need not be stipulated a UG-based set of syntactic categories (or formal features for that matter), the question rises why syntactic categories are needed in the first place. On the basis of empirical (typological) research I will try to explain the existence of syntactic categories, redundancy and doubling effects as a result from mismatching phonological and semantic economy conditions. The research will include studies of negation, modality, tense & aspect, case and Wh. The study will not only investigate cross-linguistic variation, but also focus on grammaticalisation. A final objective of this study is to account for parametric variation in terms of the variation space induced by the above-mentioned conflicting economy conditions. Consequently, such an approach enables us to consider parameters as not being innately present, but being derived as a by-product from language acquisition. |
